Approach to the Target Data Association Problem Using Subjective and Statistical Information. (Reannouncement with New Availability Information).

摘要

A general approach is presented to the data association problem in tracking which incorporates both linguistic-based and statistical information, extending classical results. The technique continues the previous effort of the PACT (Possibilistic Approach to Correlation and Tracking) algorithm. In addition, asymptotic consistency results are obtained relative to increasing sample sizes and model refinements. The PACT algorithm has been documented in previous papers. It is an attempt to perform track-to-track or report-to-track data association or 'correlation' based upon varying information sources. The procedure in an obvious way may be extended to problems of medical diagnosis, fault determination, and classification.